The Indians Of Silicon Valley The hidden geniuses of the tech revolution are Indian engineers--here's how one bucked stereotypes, got rich, and has become godfather to a generation of immigrant entrepreneurs.
By Melanie Warner

(FORTUNE Magazine) – Kanwal Rekhi's got these big eyes that flicker when he talks. His shoulders are rounded and his large, thick frame shows marks of age. But when he speaks, his face, large and jowly, emits youthful energy. His speech is rapid and subdued, perhaps racing to keep up with his mind. Rekhi often leaps into new thoughts and words before he's finished the last ones, which can make him hard to understand. This effect is compounded by a slight accent, an artifact of Rekhi's having spent most of his first 18 years in Kanpur, India. Between the speech and the physique, he has more than a passing resemblance to Vito Corleone, the Mafia don played unforgettably by Marlon Brando in The Godfather. Which is rather fitting because Rekhi, 54, is the unofficial but quite undisputed godfather of Silicon Valley's Indian mafia.

Not that it's really a mafia, though some people call it that for fun. Rekhi, who sold his company, Excelan, to Novell in 1989 for $210 million (which was, at the time, believe it or not, a lot of money), doesn't strike the slick pose Brando assumed as Don Corleone. Rekhi's gray hair is always slightly disheveled, and befitting his background as an engineer, he comes off as disarmingly rumpled. Last year, Rekhi wore a tie exactly six times. For all but a few occasions, he dresses new-economy casual--khakis and a cotton shirt that's likely to say something on it like EXODUS.COM and to have been a freebie from a company he's funded.

When people call Silicon Valley's Indian population a mafia, they mean that the immigrants who live in the Bay Area and work in high tech--roughly 200,000, according to siliconindia magazine--have formed an amazing web. Indians invest in one another's companies, sit on one another's boards, and hire each other in key jobs. Many live in close proximity and hang out together. The network might be only mildly interesting if so many of the Valley's Indian immigrants hadn't become phenomenally wealthy and successful in the past ten years. People don't necessarily think of it this way, but Bay Area Indian immigrants represent America's most successful immigrant group. Collectively, they've created companies that account for $235 billion of market value. If you add up just the net worth of the people mentioned in this story, for instance, you'll get more than $8 billion.

It's safe to say that without Indian immigrants the Valley wouldn't be what it is today. Indian engineers have been coming to the U.S. in increasing numbers since the early 1970s; almost half the H-1B visas given by the State Department go to Indian engineers (H-1Bs are granted to foreigners who have specialized skills or are, oddly enough, fashion models). High-tech companies need people desperately--U.S. engineering schools simply don't produce enough graduates to fill the specialized jobs the high-tech industry creates. According to the Information Technology Association of America, a trade group in Arlington, Va., more than 800,000 infotech jobs will go begging this year. And the engineers U.S. schools do produce typically aren't as talented as those from India. Many Indian immigrants have graduated from schools that make Harvard and MIT seem easy to get into. The six Indian Institutes of Technology, created in the 1940s by Prime Minister Jawaharlal Nehru to educate engineers for public-works projects, produce some of the world's smartest techies. Last year about 3% of the students who applied to IIT got in, while 11% to 18% were accepted at Ivy League schools.

Rekhi went to IIT Bombay and is personally worth more than $500 million. As head of The Indus Entrepreneurs, or TiE, the Valley's large, influential Indian networking group, he's at the center of a whole lot of wealth creation. Since he left his job as chief technology officer of Novell in 1995, Rekhi has poured millions of his own dollars into more than 45 startups founded by Indians, many of them IIT grads. But this sort of angel investing is not really what Rekhi considers his true calling. If you ask what he does, he'll tell you he mentors entrepreneurs. That means, among other things, that every week he spends one or two days meeting with other Indians who want to talk about the companies they've started. (Rekhi says someone doesn't have to be Indian to get a meeting with him, but most people he sees are.)

From 10 A.M. to 3 P.M., usually on Thursday or Friday, Rekhi sits in a small conference room in TiE's charmless offices in an office park in Santa Clara and listens to pitches from three or four entrepreneurs. The scene unfolds a bit like the sequence in The Godfather in which Corleone entertains a stream of requests on his daughter's wedding day. If an entrepreneur is lucky, he or she might leave TiE with the names of a few people to call from Rekhi's long list of contacts. If he is really, really lucky, he might get some combination of Rekhi's money, his commitment to join the board, and his promise to take an active role.

Because of time constraints, Rekhi does this with only a few companies at a time. Those he's actively involved in now--Ensim, Instantis, 123SignUp, and Paramark--were all founded by Indians who attended IIT. Rekhi gave several hundred thousand dollars to each, but lately he's been saying he's not going to do that anymore; he realized last fall that he had so many investments, from both his own bets and those of the half-dozen large Valley venture capital firms he's invested in, that he'd become more a business than a person. Although he's made more money by funding entrepreneurs than he did by creating a company and selling it to Novell, Rekhi insists he doesn't care about earning more. He'd rather help entrepreneurs for free. "I've got more money than anybody should. More than I can spend in ten lifetimes," he says.

At a recent TiE networking event at the Santa Clara Marriott, Rekhi was doing what he calls "making myself available." It was a Sunday afternoon and 400 people packed a large conference room to schmooze, compare notes, and, if they wished, get up on stage and make a two-minute pitch for themselves or their companies. Rekhi stood in the back of the room talking to people who opted to take their story straight to him. For about an hour an eager congregation surrounded him. Some he gave advice to on the spot. Others he told to call the TiE office and set up a meeting with him. Others he told he couldn't help at all. When he'd finally worked his way through the crowd, he'd collected 20 business cards. "There are too many. I can't keep up," he said, shuffling through them.

It wasn't always this way. Indians didn't network with one another when Rekhi and seven other midcareer Indian entrepreneurs started TiE in 1992. Going even further back, Rekhi remembers a time when hardly any Indians were starting companies. In 1982, when he founded Excelan with Subhash Bal and Inder Singh, friends he'd made while working briefly at a startup, most of the Indians he knew were working as techies at big companies. The trio's plan to run Excelan, which made gear for local area networks, was unusual because Indians were widely regarded as great techies but inadequate managers. So when three Indians who lacked a white guy went to raise money from VCs, they faced lots of slammed doors. Recalls Rekhi: "We'd always hear that the company didn't have a 'businessman,' that there wasn't anyone with a marketing background or selling expertise. That's what they'd say. But the real issue was, Will customers buy from an Indian? Indians were seen as damn good backroom operations people, but are they good in the front room, running the show and selling to customers?" This helps explain why, despite the wealth and success of Indian entrepreneurs, there are still few Indian CEOs running high-tech companies. In many instances, VCs investing in an Indian-founded company have brought in a non-Indian CEO, relegating the founder to a technical role. In other cases, large companies have acquired Indian startups for their technology or their techies. Try to name half a dozen big high-tech companies with an Indian CEO; ten bucks says you can't.

Rekhi and his partners did end up raising the money to fund Excelan, but Rekhi never got a full chance to prove himself as a boss. Now lots of small companies he has invested in are headed by Indian CEOs intent on remaining chief through an IPO and beyond. Like the dad who never made the cut for the varsity team, Rekhi delights in seeing members of a new generation of immigrants become superstars.

When he and his partners started Excelan, they decided that Singh should be the boss because he was the smoothest and most articulate. Despite his quick mind and knack for strategy, Rekhi lacked polish. He'd worked for 11 years at big companies, like mainframe-maker RCA Systems and government contractor Signal Link, but had never risen above systems engineer. Rekhi remembers asking about management jobs, then seeing people he considered less competent be promoted over him. "They'd say, 'You're a brilliant technologist, so we can't afford to lose you.' And they'd promote this other guy who didn't know as much," says Rekhi.

He got to serve as Excelan's interim CEO after Singh didn't work out. Though his term was brief, Rekhi worked hard to grow into the role. He shaved his beard and bought lots of suits and white shirts. On the advice of several directors, he underwent speech therapy for a stammer that had dogged him since childhood. To better understand the selling process, he moved his cubicle from engineering to the marketing department and often tagged along on customer calls with Excelan's sales guys. Rekhi liked being boss and thought he was good at it. But when the time came for Excelan to go public, the board decided it needed someone who would play to Wall Street; it hired Dick Moore, a 25-year Hewlett-Packard veteran. Rekhi stayed as EVP and board chairman, and most employees still thought of him as the boss. Moore did pull off a successful IPO but couldn't adapt to startup culture. Within a year he was gone. Again Rekhi stepped into the top spot until Novell acquired the company in 1989.

Looking back, John Dougery, an Excelan investor and director, admits that the board probably should have let Rekhi stay as CEO in the first place. "Back then, Indians weren't perceived as winning CEOs. We didn't know if people would trust them as managers," says Dougery.

By the time Rekhi got to Novell, he knew a lot about business and felt he was qualified to run a large public company. In the early 1990s, CEO Ray Noorda was on his way out, and Rekhi, who was CTO, made it known that he wanted the top job. The board considered him briefly, only to opt for Bob Frankenburg, another man from HP. Rekhi had watched Noorda make what he considered dumb decisions--spending more than $1.2 billion to buy the PC software packages DR/DOS and WordPerfect. ("They were dead applications," says Rekhi.) Now he watched as Novell's stock sank and the company floundered under Frankenburg. By 1995 he had decided to quit. He took a vacation of several months with his wife and two kids, then packed up his boxes at Novell and carted them to his home in Los Gatos, Calif. For the first time in his life, he had nothing to do.

At first Rekhi went into a funk. "When you become successful, you become incompetent as a person, because all the little things are done for you by others. I didn't even know how to make airline reservations or who to call to get a PC in my home," he says. To ease the boredom and quit bugging his wife, he started hanging around TiE. Though Rekhi had helped found the organization, he hadn't been active. In fact, he says that until 1995 he had little connection to the Indian community. His wife, Ann, is American (they met in Florida while he worked at a computer company and she served in the U.S. Air Force), and at least half their friends are American. Indian food is rarely served at home; Rekhi listens to Indian music only when driving alone. "I was as de-Indianized as anybody," he says. But as he hung around TiE, something started to happen. People heard that the former CTO of Novell was there and began showing up to talk. Eventually Rekhi was advising as many as eight to ten entrepreneurs a day.

On a recent morning, Rekhi is scheduled to meet with Jay Karmarkar, a Ph.D. who was in the same class at IIT. Karmarkar arrives promptly at 10 A.M. and sits opposite Rekhi in TiE's conference room. Karmarkar has spent the past decade running a small government contracting firm that does things like algorithm development, and looks the part. His glasses are one or two sizes too big, and his handshake transfers nervous energy that suggests meetings like this are a new undertaking. Karmarkar is here to see Rekhi because he has an idea for a network computer that will run on a stripped-down Unix operating system and be compatible with Linux applications. After chitchat about the genesis of his startup, Karmarkar launches into his business plan, which he's laid out neatly on sheets of paper on the table. "What we want to do is an Internet terminal that we'll sell for really cheap, probably $200. It'll be light and efficient because we're taking out all the bloated applications of the PC."

Karmarkar barely gets to page three when Rekhi interrupts. "Stop right there. Everyone who's tried to build a Web PC hasn't succeeded. Larry Ellison's lost millions on that," says Rekhi, leaning forward, elbows on the table.

"Yes, but we've got a whole new idea for the operating system," Karmarkar says, shuffling his papers in search of the slide that will respond to Rekhi's objection. Rekhi doesn't let him get to the slide. "I can get a $500 PC right now; how can you make it cheaper?"

"We're going to build a stripped-down version of Linux..." Rekhi interrupts again. "The reason PCs are successful is because you have all these applications and software inside them. If you take all that stuff out of the PC, you end up having basically a TV. You take out all the value."

Ten minutes later, Karmarkar is slumped in his chair, listening to Rekhi tell him how he'll have to convince armies of developers to design applications for his operating system. "People will look at you, this new guy, and think, 'Why should I trust him?' How can you convince them you're reliable?"

Karmarkar nods and says, "You have a point there."

One reason Rekhi is so sought after by entrepreneurs is that he's brutally honest. He can listen to someone and, within ten or 15 minutes, understand exactly what he is trying to do and offer unrestrained feedback. Unlike many VCs, who like to string along entrepreneurs in case they later have a change of heart about funding them, Rekhi calls everything as he sees it. "Totally lost" is his assessment of Karmarkar when I ask about the meeting several days later. Rekhi didn't use those exact words with Karmarkar, but when Karmarkar requested that Rekhi refer him to some investors, Rekhi replied point-blank, "I wouldn't send you to anybody until you figure out your plan. I have to feel good about someone when I refer them. Or else people call me up and say, 'Why did you send me that one?'" When I catch up with Karmarkar a few weeks later, he says he's never taken "that many bullets in ten minutes," but that a lot of what Rekhi said was right and that he's thinking about how to revise his pitch.

Most entrepreneurs Rekhi sees are savvier than Karmarkar, but many still orbit in a techie bubble. They load their presentations with details of exactly how their technology works. They use words like "pipelining" without explaining what they mean. They drone on after people have tuned out. Since most Indian immigrants come from a technical background, many find the social nuances of business daunting. "All these companies could have benefited from an hour with me," Rekhi declares, as we head out of one of TiE's monthly angel-investor meetings at the Silicon Valley Capital Club in San Jose. Five entrepreneurs have just presented their startups to a room of about 70 angel investors and not one gave a clear, compelling pitch. "That second one, I still can't understand what they were doing," says Rekhi.

By far Rekhi's most successful mentoring achievement is K.B. Chandrasekhar. Chandra, as he's known, isn't the wealthiest Indian in the Valley--that distinction goes to Vinod Khosla, a co-founder of Sun Microsystems and partner at Kleiner Perkins--but as founder of Exodus Communications, Chandra's the best-known Indian Internet entrepreneur. He pioneered server farms, those Internet data centers that house and maintain other people's Web servers. Now Exodus is a $242 million-a-year company that employs 2,000 people. But when Rekhi met Chandra in 1995, he was running Exodus out of a small office in Sunnyvale that was overheating because of all the servers stacked on tables. He was almost out of money and, if he wasn't able to raise more, would probably be returning to India with his wife and three kids. Rekhi became his savior.

They met at a TiE event. Chandra remembers the encounter better than Rekhi because Chandra was one of about three dozen people who came up to Rekhi that night. For some reason, Rekhi wrote on the back of Chandra's business card "call this guy"; he can't recall why, but he figures he must have been struck by Chandra's energy. When Rekhi called two months later, Chandra told him, "I need $200,000 for payroll and I want to talk to you today." Chandra had gone to the TiE event desperate. He and co-founder B.V. Jagadeesh had been funding Exodus from money they'd made developing software for large companies, but that wasn't going to last. Chandra didn't know any VCs or other people with money. He'd been in the U.S. just five years and, unlike most Indians in the Valley, hadn't gone to an American grad school or an IIT--Chandra's engineering degree is from the lesser-known Madras University. "My networks were pretty thin," he says.

Rekhi liked Chandra's ability to tell his story. He pegged the guy as a natural entrepreneur driven to see his ideas come to life no matter what the odds. But Exodus was an unfocused mess. Chandra was working on too many ideas. Says Rekhi: "He thought he could fund one business from the other, but I told him that almost never works." Rekhi gave Chandra the $200,000 he wanted, but told him he needed to focus on one thing and suggested it be hosting other people's Websites. Rekhi agreed to join the board and rounded up other investors, who together gave Exodus a first round of $3.2 million. Some of the investors were TiE members and others friends of Rekhi's, like construction executive Ed Shay and venture capitalist John Dougery, who have since invested in almost every one of Rekhi's deals. Helping Chandra proved Rekhi's best investment decision ever. The $1 million he eventually put into Exodus for an initial 2% stake is now worth $130 million.

Chandra says he'd like to emulate Rekhi and mentor other entrepreneurs. But for the moment he's trying to build a company he thinks will be bigger than Exodus. JamCracker, Chandra's second startup, provides an online portal where IT managers can access and manage all the software their companies need. Rekhi is an investor and a director, but this time it's not because Chandra needed money. Far from it. After turning away numerous funding offers earlier this year, Chandra raised his first chunk of capital at a rather stunning $80 million valuation. The price was too steep for Rekhi, so, as a thank-you for Exodus, Chandra sold him some of his own shares at a lower price. "I always ensure Kanwal gets a good deal. He's my guru and my mentor," says Chandra.

That sort of thing happens a lot in the Indian network. Take serial entrepreneur Naren Bakshi. When he presented Versata, his first startup, at TiE's annual conference in 1995, he was really pining for the $800,000 Rekhi and 12 other TiE members eventually gave him. For his second startup, Xpede, a financial services company, and his third, a dot-com called BuildYourDreamHome, Baskhi, 56, had VCs lining up. He still cut Rekhi and other TiE members in on the deals. He told Xpede's VCs that $1 million of the $10 million first round had to come from these angels.

As in all good networks, reciprocity keeps the wheels turning. Nimish Mehta, CEO of Impresse, which runs a Web service for corporate printing needs, was among the TiE members who invested that $800,000 in Versata. An investor in JamCracker, he's also friends with Chandra. So when Impresse was raising capital, Chandra got to do something virtually every investor in the Valley would kill to do--he invested alongside Kleiner Perkins and Benchmark. Rekhi, because he's the godfather and everyone wants him to invest, also got in on the deal. As did Satish Gupta, CEO of Cradle Technologies, a chip company. Why him? Impresse CEO Mehta, who left a job as senior vice president at Oracle, is one of four Indian angel investors in Cradle.

The wealth-creation potential of this network is huge. So naturally it was only a matter of time before the Valley's money establishment discovered it. The same VC firms that used to look warily at Indians starting companies now trip over themselves trying to get Indian deals in the door. Scott Sandell, a partner at New Enterprise Associates, asks the Indians he's already funded to send their entrepreneurial friends over to NEA. Of the 15 companies Sandell has funded at NEA, four were founded by Indians.

What VCs have realized is that Indians are not only some of the best entrepreneurs they've ever met, but they're also wired into one of the Valley's most precious resources--technical talent. Companies with Indian founders, especially ones who've gone to IIT, can often hire teams of developers faster than the average company. This can be a huge advantage in the Internet race to get up and running ahead of the other guy. Munjal Shah, a first-generation Indian American, says he's had no problem hiring top-flight engineers for his company, Andale, which provides services to people who sell lots of stuff on auction sites. Shah met his two chief engineers, both IIT grads, through one of his angel investors, all seven of whom went to IIT. Shah thinks the connections enabled Andale to move much faster than it would have otherwise. "Our time to team was lower," he says.

One recent day, as I walked through TiE's offices with Rekhi, he showed me the brochure for the group's upcoming annual conference. The first few years TiE held a conference, nobody of note wanted to come and speak. "This year we had people calling us," Rekhi said with great pride. Jim Clark and Sycamore Network's Gururaj Deshpande are the keynotes; McKinsey CEO Rajit Gupta is also speaking. Flipping open the brochure, Rekhi pointed out a long list of conference sponsors--Goldman Sachs, Morgan Stanley, Kleiner Perkins, KPMG, Internet Capital Group, and so on. The conference quickly sold out; 1,500 people signed up.

Such accomplishments attest to Rekhi's success in developing the Indian network. "Some of my friends joke that we don't need TiE anymore," he says, which is why Rekhi has begun to turn to a place he thinks needs a lot of attention--India. "I call it the economic freedom movement," he says. "Information technology is restoring India's confidence. The change is amazing." Among Rekhi's India projects is a campaign to raise private money for IIT, which, despite its wealthy graduates, is still funded by $100 million a year from the Indian government. He kicked off the campaign by giving $5 million to his alma mater, IIT Bombay. Each time he visits India, which is frequently these days, Rekhi meets with Prime Minister Atal Behari Vajpayee or cabinet members to discuss economic reform. Rekhi wants the government to foster U.S.-style venture capital, so that India will become a haven for startups and not just a place to farm out cheap software development. In the face of India's one billion people and notoriously sclerotic and corrupt bureaucracy, these would seem ambitious goals. But Rekhi is undaunted. "I have the time and the financial resources," he says. "I want to transform India." Don Corleone, it seems, is starting to act a little like Gandhi.